ProLoaF: Probabilistic load forecasting for power systems

SoftwareX, vol. 23, pp. 101487

Abstract

Today, the energy supply does not follow the demand in a controlled manner anymore. Thus, forecasting the electricity consumption became essential for the operation of power systems. Already numerous open source software tools exist that provide forecasting models, which are configurable for different forecasting tasks. In the case of electrical energy demand, a change in the geographical or temporal settings, requires specific domain knowledge on relevant data and influencing factors that are to be considered when developing data-driven forecasting models. With ProLoaF , we propose a holistic machine-learning based forecasting project, which offers the developer a continuous deployment of reliable forecasts for the power system domain. ProLoaF serves for probabilistic forecasts of the electric energy consumption and non-controllable generation in future power system operation. By overlapping Machine Learning (ML), DevOps and power systems engineering disciplines, we aim to accelerate future forecasting model development by reducing consultation work between domain experts.

Authors 2

1. Gonca Gürses-Tran corresponding

   Fraunhofer Institute for Applied Information Technology

   Affiliation as printed

   Fraunhofer-Institut für Angewandte Informationstechnik FIT, Hüttenstr. 5, Aachen, 52068, Germany

2. Antonello Monti Aachen Institute for Automation of Complex Power Systems

   RWTH Aachen University · Fraunhofer Institute for Applied Information Technology

   Affiliation as printed

   Fraunhofer-Institut für Angewandte Informationstechnik FIT, Hüttenstr. 5, Aachen, 52068, Germany

   Institute for Automation of Complex Power Systems, RWTH Aachen University, Mathieustr. 10, Aachen, 52072, Germany